Okay, reposting this answer in something that I can cut because holy shit that was longer than I thought.
Anyway, let me tell you about underpainting and color theorystuck
I do monochrome under paintings most of the time. They’re extremely helpful for lots of things, like testing compositions and planning values. It depends on how much I’ve planned the picture and how much effort I want to put in, but under-paintings can be complete value studies or just a few squiggles with rough suggestions of value. With me, most of the time color schemes are completely negotiable after I’ve finished with the monochrome underpainting. However, the color temperature is not. You see, by “monochrome” I don’t mean black-and-white, I mean what it says on the tin, just one color. That color could be blackish green (lotta historical precedent for that; it’s good for intimidating renaissance works) red, blue, whatever. What matters is the color temperature. Whatever temperature the underpainting is, that’s going to be the predominant temp and associated psychological mood of the painting. For instance, if you were going to follow Picasso into a Blue Period you’d use a blue hued under-painting with values reaching from a dark blue that’s nearly black to white or almost white. And your paintings would be sad. Sad and blue. And cool.
NOTE: Gotta emphasize this. When I say “one color”, I mean ONE. It’s easiest to get a dark color, like Payne’s Gray (which is blue) and lighten it up by adding more medium or reducing the opacity/layering strokes with a photoshop brush set to multiply to get different tones. Just don’t get a medium color and add white or black. White takes long to dry in natural media, and both have a tendency to desaturate or corrupt colors.
I’ll talk about computer art later—much later—but for a moment let’s go au naturale. When working in natural media, like oil, you HAVE to let the underpainting dry before going on to the next stage. If you’re too impatient, or if you’re working in a soluble medium such as goache or watercolors, then your paint is going to pick up the pigment already on the canvas. If you plan to build up and mix your colors in layers, or if you plan to let the underpainting show through in places you also have to be wary of the body of the paint. The ideal paint would be thinned out with medium enough to be semi-translucent, but still have enough opacity to have appreciable coverage and effect. If you plan to cover the whole underpainting, and want to do it accurately, it’s best to use higher opacity, fuller bodied paint and to do your color mixing on the palate and not on the painting itself. This is just in case you do pick up some of the pigment underneath. If your medium has too much solvent, or if your paint is naturally quite translucent itself—like Alizarin crimson—this might be a problem you’d have to look out for.
The oil medium I use for thinning out colors to glaze shit is equal parts mineral spirits and cold pressed linseed oil, btw. There are a lot of pre-mixed mediums, like liquin, but they have their own weird effects to them and many are expensive so fuck that noise.
Another common pitfall when mixing colors and coming out with desaturated or muddy colors is that you might be using colors with conflicting temperatures or tendencies on top of the underpainting, or in the actual over-painting itself. Y’see, there’s no such thing as a “pure” red, blue or yellow in pigment. There are colors that are close, but no cigar. I mean, if there was then mixing vibrant, crisp and accurate colors would be easy. Perfect red + perfect blue would = perfect purple, every time. But in real life, red and blue sometimes make a good purple, but sometimes make a chalky looking, sad purple. This is because every color has a tendency, another hue or temperature that it leans towards, and they don’t particularly like to mix. If you’re picking colors to go over your underpainting that conflict with that color, then it’s going to be very hard to maintain the veracity of your value study or underpainting.
This even applies to black underpaintings. Colors that look black in concentration still might have a slight tint when thinned out, depending on your medium and the quality of paints.
Alizarin Crimson, a common red pigment, and Cadmium Red are the same hue, but they tend towards different vertexes of the secondary color triangle (i.e. purple, green, and orange). Cadmium red is warm, closer to orange (and therefore yellow) so it would go well with colors that tend to be warm. Miking cadmium red with a yellow that tends towards orange would result in a vibrant orange. Alizarin crimson tends towards purple, a colder color, so mixing it with a warm orange-leaning color would result in a muddy, impure color.
If that’s what you’re going for, great! But if, as you imply, you have trouble with desaturated, dark or low fidelity colors then you’re going to want to start taking color temp and tendencies into consideration.
BUT WAIT, you say! What if i’m working on the computer? Surely it’s easier to maintain the purity of colors on one-a them? Well…
Working on the computer in RGB could THEORETICALLY get you close to “pure” colors because technically your work is going to be just numbers that represent light. PLUS that light is automatically mixed by your monitor—but not in the same way you’d mix pigments. Light uses the additive theory. Additive theory means your primaries are red, green and blue: R,G,B. Mix them all together and you get white. You don’t have to worry about the impurities you get in natural pigments, which are mixed subtractively. In subtractive theory primaries are going to be Red, Blue and Yellow. Mix them all together and… well, you get a mess but it’s going to be the same darkish brown mess you’ve been familiar with since kindergarten. Ideally, you’d get black as but because of the aforementioned impurities of non-digital pigments that doesn’t quite happen.
Even if you use RGB in your computer, the process inks (Cyan, Magenta, Yellow and Black) use to print print images are not “pure”. Also, when you use the purest subtractive primary colors you can get from an RGB profile (which are #FF0000, #FFFF00, #0000FF) you can STILL have variable tendencies on different computers. It really depends on your monitor’s calibration and how old it is!
calibr—what was I talking about?
So anyway, my point is that in photoshop you can’t rely on what you’ve learned from working with real pigments too much because it won’t literally translate exactly. But if you do find you’re having troubles, it’s much easier to fix them using the computers. You can save undersaturated colors by manipulating them with the sponge tool (which changes saturation) or by fucking around in the levels pane.
When I do digital art I like to work in layers, same as I work in oil paints. By building up colors as you go and mixing them on the canvas itself if you can’t find the perfect shade in the color picker. THIS CAN SAVE SO MUCH TIME u wood not beleave. If your underpainting is well developed and if you want to get really robust, vibrant colors when you glaze or paint over it, go mess around on your layer palate with different layer modes. Color dodge is useful for adding a layer of brilliant, bright highlights, while Color Burn often brings out intense, dark shades but at the same time both of these modes are very dramatic and it’s easy to burn out detail. A lot of people just stick with using a multiply layer, but that doesn’t always get the most dramatic results.
….I probably just typed out like several billion words with only one relevant paragraph, didn’t I? Anyway.
[what follows in the next two paragraphs is a quick and dirty, completely extraneous explanation of why light/pigment mixing is called additive/subtractive if you don’t recognize those terms. You are excused from the rest of this answer if you don’t care.] In additive theory you get colors by combining the select wavelengths of the spectrum of visible light, in other words by ADDING colored light. Additively speaking, yellow is made of red and green, while white is just a combination of all colors in the spectrum. In RGB hex it’s FFFFFF. FFFFFF means EVERYTHING IS MAX. In other words, you’re getting all the red, green and blue you could manage all comin’ atcha eyeballs at LIGHT SPEED. And so, white! In subtractive theory, which is how we’re used to thinking about color, you see only the light that is not absorbed by the object. So when you add a primary to a primary, you’re increasing the amount of light absorbed and subtracting the light reflected. In a world of absolutes, this would mean that when you mix pure pigments of all three primaries it’d reflect no light, giving you black. In hex you describe it with zeroes, 000000, no bit of anything is getting out of that color. But we see how often that actually happens in practice.
Here’s a fun example of the shit color and light gets up to: In a red line drawing, the red ink would absorb yellow and blue and only reflect red. But if you were glasses tinted the same shade, it’d be pretty damn hard to tell though since everything the fuck would look red. The red lenses cancel out that red reflected light, while making everything else look redder. Meanwhile, if you have an exact copy of the drawing overlapping it, it would absorb red and yellow and reflect blue. If you popped out one of the lenses of your red glasses and put a blue lens in you’d not see the blue drawing out of that eye but you’d see the red. There are two images on paper, but because of the color filters you’re wearing each eye only sees one. If you have stereoscopic vision (if you’ve two eyes, it’s likely), then your mind will automatically combine each slightly different image from your two eyes into one composite image. What do you get then?
Er, wait. You get the illusion of a three dimensional object from a 2 dimensional drawing. NOW YOU TOO CAN CHARGE LIKE 20 DOLLARS A TICKET TO SEE YOUR DRAWINGS IN THREEEEEE DEEEEE!!!!!
of course that’s not how they do 3d anymore but shush
Annnnnnd none of this is is anything but tangentially related to the ask right now. THAT’S MY CUE TO STOP TYPING